REAL Expert Advisor backtesting (99% modeling quality) | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

How to achieve 99% modeling quality for the expert adviser 1 reply

M1 timeframe and modeling quality... 3 replies

Backtesting 90% modeling quality? 1 reply

How to achieve 99% Modeling Quality 2 replies

How to improve modeling quality? 6 replies

Attachments: REAL Expert Advisor backtesting (99% modeling quality)
Exit Attachments

REAL Expert Advisor backtesting (99% modeling quality)

  • #1
  • First Post: Aug 20, 2018 2:01pm Aug 20, 2018 2:01pm
As a matter of fact, the VAST majority of EAs (expert advisers) will COMPLETELY FAIL when it comes to live trading.
Even after a "3 months positive run", the probability is VERY high that thereafter, the EA will simply "die".

Some of the REASONS are related to the nature of market manipulation, which I have exposed already (see my thread on that here:
https://www.forexfactory.com/showthread.php?t=768090


Funny enough, I can already demonstrate the facts above to you, this just by running any EA over REAL TICK DATA (over some years), this also incorporating:

- variable spread
- slippage
- commissions
- intra-bar moves
etc.


Since I DO possess the hard- and software to do respective tests with a 99% modeling quality, I hereby offer ANYONE to post or send me an EA, I will then post the respective REPRESENTATIVE test results here, this for everyone to see (you automatically give consent to that by posting or sending).

As you will see, the results will show downward-pointing account curves (some sooner than others).
If you want hints with regard to how to incorporate the actual market reality, then please read my "FX Exposed" thread.


Note:

PLEASE do NOT post / send any EAs that are based on:
- martingale
- grid
- pyramiding
- hedging
- recovery mechanisms
- no stop loss
- News

All the best & stay safe.

FF
www.sme-fx.com
  • #2
  • Aug 21, 2018 2:33pm Aug 21, 2018 2:33pm
Note:

Against what many broker trolls try to convince you of, because of the MT4 GENERIC TICK SIMULATION, any outputs from the strategy tester without additional software is COMPLETELY INCORRECT.
www.sme-fx.com
  • #3
  • Aug 21, 2018 3:12pm Aug 21, 2018 3:12pm
https://www.mql5.com/en/articles/1511
www.sme-fx.com
  • #4
  • Aug 21, 2018 3:13pm Aug 21, 2018 3:13pm
Attached Image (click to enlarge)
Click to Enlarge Name: wswswd.png Size: 40 KB
www.sme-fx.com
  • #5
  • Sep 5, 2018 8:19pm Sep 5, 2018 8:19pm
Hi FF if you ever have some time to kill could you run a test on this EA? Much appreciated mate.

https://www.mql5.com/en/market/product/30134#
  • #6
  • Sep 16, 2018 12:59pm Sep 16, 2018 12:59pm
Ran it:

- tickdata
- variable spread
- slippage
- commissions


EU M1, over 2015........VERY few trades:
Attached Image (click to enlarge)
Click to Enlarge Name: sfewfewfewfew.png Size: 38 KB
www.sme-fx.com
  • #7
  • Sep 16, 2018 1:01pm Sep 16, 2018 1:01pm
2015, EU, M1:
Attached Image (click to enlarge)
Click to Enlarge Name: wedewefewf.png Size: 16 KB
www.sme-fx.com
  • #8
  • Sep 16, 2018 1:01pm Sep 16, 2018 1:01pm
Attached Image (click to enlarge)
Click to Enlarge Name: ewfewfewfwedewdew.png Size: 23 KB
www.sme-fx.com
  • #9
  • Sep 16, 2018 1:03pm Sep 16, 2018 1:03pm
I show you soon a BT with tickdata compared to without....
www.sme-fx.com
  • #10
  • Edited 11:28pm Sep 16, 2018 1:10pm | Edited 11:28pm
Quoting FractalFreak
Ran it: - tickdata - variable spread - slippage - commissions EU M1, over 2015........VERY few trades: {image}
Can you test an EA with does the following:

  1. Multi-Pair either 28 or 72. List of trade ablpairs change every hour multiple times a day
  2. Multiple timeframe...more than the standard 9 TFs we see on our charts (automatically chosen and changing on every minute candle close)
  3. Multiple strategies (automatically chosen and changing on every tick)
  4. Time based filters-backtesting MUSt take into account local and server based DST with 100% accuracy throughout the backtest range.
  5. Changing of over 300 parameters on every tick.
  6. Entry setups could be over 450K (i.e. unique magic numbers)
  7. And many other features.


It .will take over a month to run a backtest of 5 yrs data :-)

Staying in my lane...
  • #11
  • Sep 16, 2018 2:42pm Sep 16, 2018 2:42pm
Aw thank you FF, your the man. )

Cheers!
  • #12
  • Sep 17, 2018 2:22am Sep 17, 2018 2:22am
Quoting VEEFX
{quote} Can you test an EA with does the following: Multi-Pair either 28 or 72. List of trade ablpairs change every hour multiple times a day Multiple timeframe...more than the standard 9 TFs we see on our charts (automatically chosen and changing on every minute candle close) Multiple strategies (automatically chosen and changing on every tick) Time based filters-backtesting MUSt take into account local and server based DST with 100% accuracy throughout the backtest range. Changing of over 300 parameters on every tick. Entry setups could be over...
Why not take the code = buy sell methods and run it on a single pair EA?
Vucking good EA coder... https://t.ly/AZjRM
  • #13
  • Sep 17, 2018 2:45am Sep 17, 2018 2:45am
Quoting reteid2222
{quote} Why not take the code = buy sell methods and run it on a single pair EA?
Because a diversified multipair, multi-TF, multi-session, multi-strategy EA will almost always fail when run on a single pair EA. Why? Because there is no pair or TF specific optimization of parameters involved to make the backtesting curve look pretty :-
Staying in my lane...
  • #14
  • Sep 17, 2018 7:04am Sep 17, 2018 7:04am
Quoting VEEFX
{quote} Because a diversified multipair, multi-TF, multi-session, multi-strategy EA will almost always fail when run on a single pair EA. Why? Because there is no pair or TF specific optimization of parameters involved to make the backtesting curve look pretty :-
of course
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 398 KB
Vucking good EA coder... https://t.ly/AZjRM
  • #15
  • Sep 17, 2018 5:19pm Sep 17, 2018 5:19pm
hey what do you think? id like to see your sophistiated backtest report.

based on this thread
https://www.forexfactory.com/showthread.php?t=635569
Attached File(s)
File Type: ex4 xm7_Ccfp_Wkly_TradeMonitor_v40.4.ex4   215 KB | 480 downloads
Profit growth tells you the trader ability.
  • #16
  • Sep 17, 2018 11:30pm Sep 17, 2018 11:30pm
Quoting voketexpert
hey what do you think? id like to see your sophistiated backtest report. based on this thread https://www.forexfactory.com/showthread.php?t=635569 {file}
I m interested HOW this is possible to test if someone proclaims this...
I would simply do it as indicator=calculator if someone=voketexpert is interested....
I m not interested in tradining or coding systems which run some weeks and see the bad results...
Vucking good EA coder... https://t.ly/AZjRM
  • #17
  • Sep 18, 2018 7:13am Sep 18, 2018 7:13am
"
hey what do you think? id like to see your sophistiated backtest report.
based on this thread"


Will do it for you soon....

I am running some other stuff at the moment.......will post soon..

best, FF
www.sme-fx.com
  • #18
  • Sep 18, 2018 9:58am Sep 18, 2018 9:58am
Any EA here that is tested yet?
Vucking good EA coder... https://t.ly/AZjRM
  • #19
  • Sep 18, 2018 10:05am Sep 18, 2018 10:05am
"Any EA here that is tested yet?"


Yes.

Here is one I made.

NOTE:

- Do not ask me to sell it to anyone, because I do not sell it.
- You can see that it IS possible, so do not give up.
- It took me a few years to create this


Here you go (2% risk per trade):
www.sme-fx.com
  • #20
  • Sep 18, 2018 10:06am Sep 18, 2018 10:06am
Attached Image (click to enlarge)
Click to Enlarge Name: pf10 w f.png Size: 45 KB
Attached Image (click to enlarge)
Click to Enlarge Name: pf w f.png Size: 19 KB
www.sme-fx.com
0 traders viewing now